Replication data and code for: Who Bears the Downwind Burden of Cooling? Coal Power, PM2.5 Exposure, and Energy Justice in China

Description

This dataset contains the replication data and code for the manuscript "Who Bears the Downwind Burden of Cooling? Coal Power, PM2.5 Exposure, and Energy Justice in China." It includes analysis-ready province-day, receptor-day, source-receptor-day, source-receptor-year, electricity, vulnerability, and geometry panels for 2017--2020; derived result files underlying the tables and figures; scripts for producing the empirical outputs; metadata; and SHA-256 file manifests. The original third-party source layers are documented in dataset_catalog.csv and the data guide. No synthetic or manually fabricated observations are included. Reuse of materials derived from third-party sources remains subject to the terms of the original providers.

Steps to reproduce

1.Regenerate Empirical Outputs Run the scripts in this order: python scripts/40_build_erss_paper_outputs.py python scripts/45_build_review_revision_data_supplements.py python scripts/50_build_referee_revision_outputs.py python scripts/55_build_second_round_referee_outputs.py python scripts/46_build_sixth_round_robustness.py The order matters. Later scripts use intermediate files or result files created by earlier scripts. Script roles: 40_build_erss_paper_outputs.py builds the core heat-emissions estimates, main downwind PM2.5 exposure estimates, timing diagnostics, inference checks, heterogeneity outputs, context maps, and source-receptor accounting outputs. 45_build_review_revision_data_supplements.py rebuilds supplementary processed files for pollutant, vulnerability, and incidence analyses. 50_build_referee_revision_outputs.py builds source-heat reduced-form checks, pollutant triangulation, vulnerability tables, and identifying variation outputs. 55_build_second_round_referee_outputs.py builds the source-heat audit, leave-one-source-out checks, source concentration diagnostics, FDR-adjusted tables, and related figures. 46_build_sixth_round_robustness.py builds transport-accounting ranking, vulnerability, and pre-2020 robustness outputs. Main generated folders: outputs/erss_paper/data/ outputs/erss_paper/tables/ outputs/erss_paper/figures/ processed/revision_supplements/ processed/sixth_round_revision/ Expected non-fatal warnings include absorbed fixed-effect warnings, package future warnings, and LaTeX underfull hbox warnings. Missing input files, missing Python packages, or empty output folders should be treated as errors. 2.Key Outputs To Check Summary files: outputs/erss_paper/data/analysis_summary.json outputs/erss_paper/data/referee_revision_summary.json outputs/erss_paper/data/second_round_revision_summary.json outputs/erss_paper/data/sixth_round_revision_summary.json outputs/erss_paper/data/seventh_round_directional_magnitude_summary.json Important CSV outputs: outputs/erss_paper/data/downwind_robustness_coefficients.csv outputs/erss_paper/data/downwind_lead_lag_coefficients.csv outputs/erss_paper/data/downwind_directional_diagnostics.csv outputs/erss_paper/data/temperature_bin_coefficients.csv outputs/erss_paper/data/heterogeneity_coefficients_with_fdr.csv outputs/erss_paper/data/source_heat_second_round_coefficients.csv outputs/erss_paper/data/source_receptor_burden_decomposition.csv 3.Expected Numerical Checks A successful reproduction should recover these headline values: CDD24 +1 SD is associated with 7.47 percent higher coal PM2.5 emissions. CDD24 +1 SD is associated with 10.81 percent higher coal CO2 emissions. The preferred wind-directed coal-emissions index +1 SD predicts 5.57 percent higher receptor PM2.5. The preferred raw transport index has mean 224.85 Mg/day, SD 146.40 Mg/day, and IQR 203.73 Mg/day.

Institutions

• Peking University

  Beijing, Beijing

Categories

Licence